The Product Launch Company

Refine the supplied problem hypothesis into a decision brief covering who might benefit, what the local demo actually does, what it excludes, and what would disconfirm usefulness. Do not claim user validation.

Inputs: starter/docs/product-brief.md, starter/data/sample-agenda.json, starter/docs/launch-checklist.md

Outputs: deliverables/product-boundary.md

Depends on: No prerequisites

• Time budgeting is separate from calendar scheduling and meeting facilitation.
• The brief preserves the synthetic nature of the sample.
• A measurable disconfirmation criterion and owner approval boundary are included.

Inspect the local demo and design spec. Create a walkthrough for loading the sample, editing topic minutes and required flags, recovering from a blocked plan, importing a local JSON file, and exporting an explicitly requested result.

Inputs: deliverables/product-boundary.md, starter/docs/design-spec.md, starter/docs/test-protocol.md, starter/demo/index.html, starter/demo/app.js, starter/demo/styles.css

Outputs: deliverables/interaction-walkthrough.md

Depends on: product-boundary

• The walkthrough can be completed without a server or account.
• Keyboard and error-recovery observations have explicit pass/fail criteria.
• Unperformed manual checks remain labeled not observed.

Run the authored Node standard-library tests and JavaScript syntax checks. Compare every structured case against the core result and record commands and observed results rather than asserting a production launch.

Inputs: deliverables/interaction-walkthrough.md, starter/demo/core.js, starter/demo/app.js, starter/reference/test_agenda.cjs, starter/data/acceptance-cases.json, starter/data/sample-agenda.json, starter/docs/test-protocol.md

Outputs: deliverables/numerical-evidence.json

Depends on: interaction-plan

• All structured numerical cases pass, including blocked capacity and midnight bounds.
• The pure core does not mutate imported input.
• Automated results do not stand in for browser or assistive-technology observations.

Adapt the supplied landing copy to the agreed product boundary. Clearly distinguish the usable reference demo from planned polish, validation, and public availability.

Inputs: deliverables/product-boundary.md, starter/docs/landing-copy.md, starter/docs/product-brief.md, starter/demo/index.html

Outputs: deliverables/launch-copy.md

Depends on: product-boundary

• Headline, short description, three feature statements, limitations, and CTA agree with the demo.
• No claim of customers, time saved, certification, AI behavior, or a generated video is invented.
• The CTA is an owner-reviewed local-demo invitation, not an unapproved external post.

Turn the support playbook and observed numerical behavior into a quickstart, troubleshooting flow, and issue-intake fields. Do not request private meeting content by default.

Inputs: deliverables/numerical-evidence.json, deliverables/launch-copy.md, starter/docs/support-playbook.md, starter/data/sample-agenda.json

Outputs: deliverables/support-kit.md

Depends on: numerical-verification, launch-copy

• Instructions explain blocked plans, unscheduled optional topics, and explicit local export.
• The intake requests reproduction with synthetic topics before personal content.
• No retention, support response SLA, or staffed service is promised without an owner decision.

Select proposed distribution surfaces and describe exact content, audience, publication owner, success measurement, and takedown route. Produce a request only; do not publish, spend, collect analytics, or contact people.

Inputs: deliverables/product-boundary.md, deliverables/launch-copy.md, starter/docs/launch-checklist.md

Outputs: deliverables/channel-approval-request.md

Depends on: launch-copy

• Every proposed external effect has an explicit approval gate.
• A local zip or static-file delivery is distinguished from a hosted service.
• Tracking is off by default and no distribution outcome is fabricated.

Perform the manual protocol if an authorized local browser is available. Record each actual observation; otherwise record not observed and keep the release gate conditional. Never infer screen-reader behavior from unit tests.

Inputs: deliverables/numerical-evidence.json, deliverables/interaction-walkthrough.md, starter/docs/test-protocol.md, starter/demo/index.html, starter/demo/app.js, starter/demo/styles.css

Outputs: deliverables/accessibility-observations.csv

Depends on: numerical-verification

• Each protocol item records environment and observed pass/fail/not-observed.
• Blocked-plan messaging and keyboard-only edit/import/export are covered.
• No accessibility certification or unperformed user test is claimed.

Review numerical tests, support issues, and accessibility observations. Produce a candidate manifest listing exact local source paths, content hashes, unresolved defects, and release blockers; do not hide not-observed manual gates.

Inputs: deliverables/numerical-evidence.json, deliverables/accessibility-observations.csv, deliverables/support-kit.md, starter/demo/index.html, starter/demo/styles.css, starter/demo/core.js, starter/demo/app.js, starter/docs/launch-checklist.md

Outputs: deliverables/demo-candidate.json

Depends on: accessibility-walkthrough, support-readiness

• All four demo files are identified with locally computed public-source hashes.
• No unresolved or unobserved gate is represented as passed.
• The candidate includes no credentials, remote dependencies, or invented signatures.

Adapt the original storyboard to the actual candidate behavior. Supply timing, captions, screen states, and truthful narration copy for later approved production; do not call a storyboard a rendered asset.

Inputs: deliverables/demo-candidate.json, deliverables/launch-copy.md, starter/docs/storyboard.md

Outputs: deliverables/video-production-brief.md

Depends on: demo-candidate

• The proposed sequence totals 24 seconds.
• Every product behavior shown exists in the candidate or is explicitly marked proposed.
• The handoff says no video, voiceover, or public distribution has been generated by this task.

Evaluate the candidate, support kit, observed interaction evidence, and channel request. Write a decision recommendation with outstanding owner approvals and a no-go path.

Inputs: deliverables/demo-candidate.json, deliverables/support-kit.md, deliverables/accessibility-observations.csv, deliverables/channel-approval-request.md, starter/docs/launch-checklist.md

Outputs: deliverables/launch-decision.json

Depends on: demo-candidate, channel-approval

• The recommendation names numerical, manual, support, and publication gates separately.
• No public launch is claimed from a local test or decision draft.
• An unresolved mandatory gate produces no-go or conditional status, never unconditional approval.

Create a sequenced runbook for later owner-approved delivery, verification, and withdrawal. Treat any video production as pending until an actual rendered file is verified.

Inputs: deliverables/launch-decision.json, deliverables/channel-approval-request.md, deliverables/video-production-brief.md, deliverables/demo-candidate.json, starter/docs/launch-checklist.md

Outputs: deliverables/launch-runbook.md

Depends on: launch-decision, storyboard-handoff

• Each publication step requires the appropriate owner's explicit approval.
• The runbook has verification and takedown steps for every proposed surface.
• Storyboard, rendered video, published demo, and observed usage are separate evidence states.

Draft questions to learn whether people understand required versus optional time budgeting. Use synthetic examples and do not collect responses or personal agendas as part of this seed.

Inputs: deliverables/launch-runbook.md, deliverables/support-kit.md, deliverables/product-boundary.md, starter/docs/support-playbook.md

Outputs: deliverables/feedback-instrument.csv

Depends on: launch-runbook

• Questions separate comprehension, task completion, and usefulness.
• Each field has a purpose and optionality; no personal agenda text is required.
• Responses and customer validation remain absent until an authorized study occurs.